Stuart Russell : 3 principles for creating safer AI

TED 2017

Stuart Russell  : 3 principles for creating safer AI (transcript)

Summary

Now, we will probably satisfy our technological development of AI. It is progressing faster than we expect. And then machines will read everything that is written by a human soon and they'll be able to make better decisions in the real world than us.

However, it will lead to the catastrophe that we can't control the robotic. To continue defensively pursuing an objective that is not aligned with the true objective of the human race but this is dangerous that it possibly  kills people to achieve one objective.

Thus AI has to have three principles. It has altruism, humility and it is useful to us.

Having altruism is to maximize the realization of human objective, but having humility is not known what its human values are. It means to avoid the problem of single-minded pursuit of an objective but to have some ideas of what we want.

In the near future, AI can also learn from what it wasn't right for human objectives and what we really want by watching all of us.

The speaker thinks that he wants to create human- compatible AI that has them.

I think that this is to have something over the human.

This must seek that we have to be better people and learn to be better people.

Words in this story

bribe /noun/ a sum of money or other inducement offered or given in this way.

malfunction /noun/  crash, breakdown, fault, failure, bug, a failure to function in a normal or satisfactory manner.